High bit-rate data signals are typically processed prior to data extraction in order to improve data recovery accuracy. While data signals can be processed to remove some types of signal defects, such processing tends to be difficult for data signals at data rates greater than a few Gb/s. In addition, signal processing circuits used to improve apparent signal quality frequently introduce new signal defects. For example, data recovery operations for signals based on the Synchronous Optical Network (SONET) frequently use so-called Gaussian or Bessel-Thompson low pass filters to limit data signal bandwidth. Unfortunately, these filters are typically reflective filters that reject high frequency components by reflection while transmitting other frequency components. The reflected frequency components can produce undesirable signal artifacts so that such filters are used with attenuators that generally attenuate all frequency components. As a result of such filtering, undesirable high frequency signal components are removed but with an overall reduction in signal level.
Establishing a preferred filter configuration for a particular data transmitter or receiver generally involves a test procedure using several filters. Based on signal quality measurements associated with these filters, a preferred filter configuration is selected and implemented. This procedure can be expensive and time consuming. In addition, measurements based on a few filters may be inadequate to identify a filter configuration that produces optimum data quality. Changes in signal source, receiver characteristics, or transmission path can be compensated only by repeating the test procedure.
Limiters or clippers are also used with high frequency electrical signals such as high bit-rate data signals. These limiters and clippers are generally based on diodes that are configured to limit signal amplitude. Unfortunately, the capacitance associated with limiter/clippers diodes degrades limiter/clipper performance, especially at high frequencies.
In view of these and other shortcomings, improved signal processing apparatus and methods are needed.